Video Analytics: Seeing Past the Hollywood Hype
By Dr. Alan Lipton
We’ve all been in the security room and seen the cameras zoom in on the bad guy, the screens on the monitor wall light up with a dramatic alert, and the heroes swoop in to save the day. Unfortunately, as most security professionals know, the real world differs quite a bit from the ultra-slick scenes in shows like Las Vegas and 24. The television and movie industries have led us to believe that surveillance technology —specifically video analytics — has capabilities that far outstrip reality. But while there isn’t a piece of software in existence that can identify a person by spotting a freckle on the back of his or her ear, video analytics has the potential to be a transforming technology that enables casinos to get much more from their existing surveillance infrastructures. But in order to derive real-world value from this new technology, it is important to understand its limitations as well as it capabilities.
Video analytics, also known as intelligent video, is software that uses artificially intelligent technology to “watch” video, extract useful information and create alerts. It is the kind of technology that fuels Hollywood’s imagination, often being portrayed as so cutting-edge that it seems more like science fiction than reality. While this Hollywood exposure has led to an increased interest in video analytics, it has also fueled unrealistic expectations for the technology, promoting a misplaced interest in far-fetched concepts and applications for catching and tracking cheats. This is happening across industries as buyers are having a difficult time cutting through the hype to learn the true capabilities — and limitations — of intelligent video. Here we’ll bridge the gap between fiction and reality in an effort to help users make better informed buying decisions.
Using video analytics for gaming table surveillance to detect cheats and other suspicious behavior is commonly misperceived. It’s expected that security and surveillance directors push the envelope and force the industry to come up with new technologies to solve their problems, but there has to be an understanding of what can and can’t be done right now with intelligent video.
So let’s first understand the myths in order to appreciate the reality. For instance, one common myth about video analytics is that it can spot a terrorist in a stadium full of people. Or that it can automatically “see” a person cheating at a Blackjack table when highly trained surveillance personnel cannot.
Myth: Video Analytics is Effective Enough to Replace Humans
When confronted with an unlikely claim about video analytics, a good rule of thumb is to ask the following question: Could a person watching the video perform the same task? If the answer is no, then it is unlikely that video analytics can do it either. Even if the answer is yes, it is still possible that the abilities of video analytics will be stretched — remember, humans have been interpreting visual images for about 5 million years. Computers have only been at it for about 40 years.
Myth: Video Motion Detection is ‘Intelligent’ Technology
The word “intelligent” is everywhere — on cameras, software packages, encoders and more. But how can a particular device be intelligent? Some include motion detection technology and some have capabilities based on computer vision science. These have widely different applications, but are often packaged under the same label. It is important to pick the devices that have the “intelligence” to meet your needs.
So it follows that state-of-the-art of video analytics technology simply cannot monitor all player movements and detect all behaviors that surveillance directors may want, especially considering the typical crowded environment around a table. For example, what’s the likelihood that all hand motions of every player will be visible? How can the analytics software know when no more bets are to be placed in order to separate the normal betting motion from illegal behavior such as past-posting of bets? Conceivably, high-definition monitoring of individual player betting areas could detect suspicious chip movement, but this would require expensive and potentially infeasible video coverage for a single-purpose piece of software.
So, What Works?
The advances being made by video analytics technology companies to detect more and more specific behaviors are significant, and there are actually many applications within a casino where video analytics are being used today to deliver immediate value and a more significant return on the investment made in video infrastructure.
More and more, casinos are making a strategic investment in video analytics to realize compelling results by deploying analytics for multiple groups within the organization, ranging from security to surveillance to marketing.
Maximizing Business Within Business
Most casinos are self-contained mini-cities providing food, retail and other services to keep guests happy, comfortable and in the casino. Having better insight into the patronage of these operations enables the casino to get an accurate perspective of customer experiences and service that provides the intelligence necessary to continuously improve operations. By using video analytics to monitor ongoing activity within individual operations, managers can make real-time decisions and assess data over time to interpret trends. For example, foot traffic for each retail store can be captured to generate threshold alerts when occupancy reaches a specific level, detecting unsafe or suspicious crowding that requires response. Or it can provide insight into occupancy trends over time, allowing managers to more accurately plan staffing levels.
For any casino, maximizing business often means maximizing customer service. Poor customer response caused by extensive lines and wait times must be monitored and corrected as it happens. Customer queues provide a particular challenge. Lines form at hotel check-in, taxi queues, ATMs, retail points of sale, show ticket booths, guest services, business center service desks and, of course, the buffet. Video analytics can monitor any queue in real time to determine the number of people in line and generate an alert when a certain level of occupancy is exceeded so appropriate action can be taken. Queue length or occupancy rules activated on video cameras throughout a casino provide a proactive and centralized method to resolve customer service issues before they escalate.
Untapped Intelligence
Casinos have invested millions of dollars in extensive video coverage throughout the entire casino property, with high-density coverage on the gaming floor. While these video systems are primarily used for both proactive and reactive surveillance, each camera represents an ongoing stream of business information that is generally untapped without analytics.
Consequently, people-counting and traffic pattern analysis are the hottest topics today for casinos evaluating video analytics. Existing cameras that provide general area surveillance on the gaming floor represent significant sources of invaluable data on how patrons move in, out and through the property. With little effort and by leveraging existing camera feeds, casinos have already started to use video analytics for this purpose. One of ObjectVideo’s customers is using intelligent video to monitor all entry and exit points on the casino floor and then aggregating that data to reliably determine the occupancy level within the casino at any point in time. That occupancy data is compared in real time to game and slot usage and to verify that staffing levels are appropriate for the current casino patron population. This same occupancy information is evaluated by the casino to determine the effectiveness of entertainers and performances in attracting customers to the casino and to measure the extent to which those customers remain in the casino once the performance has concluded.
On the casino floor, people-counting technology can also be used to measure traffic flow at any location. Casino floors represent a series of traffic paths that can be segmented and individually monitored for people flow. A great example of video analytics’ use is slot traffic, say for high jackpot slots or new machines that have just been placed on the floor. Casinos already know how often each machine is used and how long players may play the same machine. Intelligent video adds value by reporting on how many people walk by specific machines, how many people loiter in front of machines but don’t play, etc. By collecting and reporting this data over time, the casino can better assess the total traffic and conversion of that traffic to actual slot play, thereby determining the true effectiveness of the machine with its corresponding floor placement.
Count ’Em Up, Move ’Em In
Also on the casino floor, existing video cameras are being leveraged with accurate people-counting software to yield business intelligence on traffic flow, gaming usage and promotional effectiveness. Off the casino floor, there are many high-value uses for video analytics to monitor retail, entertainment and other back room operations for additional business intelligence and improved customer service. And video analytics for physical security are, of course, being applied at many points around and within the casino/resort physical environment.
And it All Started with Security
Inside any casino there are numerous restricted areas being monitored by passive video systems that can immediately benefit from intelligent video capabilities. Security personnel can spend more time on known high-risk areas while analytics picks up the continuous monitoring of all other areas, such as stairwells, fire exits, loading docks and any perimeter access point. Analytics can also accurately monitor access to and from controlled areas and compare those events in real-time with access control events to deliver a solution for tailgating and other relevant behaviors.
Clearly, the interest in video surveillance all started with security. But finding new ways to leverage the massive amounts of video coverage in casinos might just provide the edge every property is looking for in today’s competitive world. Monitoring staffing levels or evaluating patterns in floor traffic may not capture Hollywood’s imagination, but it is certainly getting the attention of forward-thinking casino managers. Video analytics may not spot every cheater who walks through the doors, but it can help optimize the customer experience, thereby maximizing each and every stay. And at the end of the day, isn’t that how the game is won?
Dr. Alan Lipton is the Chief Technology Officer at ObjectVideo, where he oversees a team of over 50 scientists and engineers that develop and commercialize computer vision technologies for applications such as security, surveillance and business intelligence gathering. Prior to working at ObjectVideo, Dr. Lipton was a member of the research faculty at Carnegie Mellon University's Robotics Institute in Pittsburgh, Pa., where he acted as a project co-manager of DARPA's VSAM project.